Co-Founder @ Syncsolutions.ai · AI/ML Engineer in progress
Building AI automation systems out of New York City.
Most RAG demos answer questions over a pile of PDFs. Enterprises can't ship that — retrieval leaks data. This project enforces access at the database level via Postgres Row-Level Security, so a buggy app query physically cannot return a chunk the caller isn't cleared for. Paired with a LangGraph supervisor, layered ShieldGemma guardrails, and a LoRA-fine-tuned Gemma 4 running fully local on Apple Silicon.
Stack: FastAPI · LangGraph · pgvector + RLS · BGE-M3 · MLX / Ollama · vLLM (Modal) · ShieldGemma · LangFuse · Next.js · Docker · OpenRouter · Tavily
Languages
ML / AI Frameworks
LLMs & Inference
Web & Backend
Cloud & Infra
Observability & Tools
- ✅ NVIDIA NCA-GENL — Certified Associate, Generative AI LLMs (2026)
- 🔄 NVIDIA NCP — Certified Professional, Generative AI LLMs (in progress)
- 🔄 Google Cloud PMLE — Professional ML Engineer (in progress)
- 🔄 Azure AI-102 — Designing & Implementing AI Solutions (in progress)
- 🔄 AWS Certified ML Specialty (in progress)
- 🔄 TensorFlow Developer Certificate — Google (in progress)
- 🔄 DeepLearning.AI MLOps Specialization (in progress)
- 🔄 DeepLearning.AI LLMOps (in progress)
- 🍎 Apple M5 Max · 64GB · MacBook Pro 16" (I much prefer this, apple on top)
- 🖥️ RTX 5090 · i7-14700KF · 256GB DDR5 (local training + Ollama inference)
